- By Andrew Garrett
- 0 Comments
Yesterday, I wrote about what it would take to get the voter verification side of things lightly computerised using the recent federal election as an example, and came up with a total spend of just over $100,000,000. Yes, One Hundred Million Dollars. And that still left us using pencil and paper in the voting booths!
So, let’s see what it would take to rectify that, dragging our electoral process kicking and screaming into the new millennium.
While I didn’t count on the day, where I voted, it looked like there were somewhere around 12 voting booths. Let’s round that down to 10, just to make the numbers easy, and make sure the final numbers will be even more conservative.
Each voting booth would need some sort of device to enable voting. I’m thinking something with a touch screen would be best, even though that adds expense. Unfortunately, we can’t assume universal computer literacy, which means that a keyboard and mouse combination would be untenable. These devices would need to be reasonably robust, somewhat ruggedised to prevent some of the less savoury elements of our population from damaging them, and secured for similar reasons. Without exhaustive research, I’m going to assign a fairly arbitrary $2,000 minimum for these devices.
With 10 booths in each of 8,642 locations, we’d be looking at $172,840,000 for the devices alone.
Oh… we forgot spares, let’s allow for another two per location, that adds another $34,568,000.
Unless we want to be shipping all of those devices back to a central location for data offload, we’ll need some sort of centralised data storage per location – it can be less physically secure, but will still need to be reliable and have a reasonable level of robustness. Let’s guess that at $4000 per unit, with one spare per every 10 locations – $38,024,800 for those.
We’d need cabling for these as well, but as they can be clustered, let’s allocate that the same amount as yesterday – another $2,592,600.
The staff at each location would need some training in how to set up, tear down, and replace these devices (in case of failure). We’d bundle that in with the other training, and allocate another $100 per person for training. Not everyone would need it, let’s say 2 people per location, that’s $1,728,400. But, we’d also need a support desk of some sort, which would have to be staffed and trained up to provide second and third level support – let’s say another $2,000,000 for that.
Now, the voting software won’t be off-the-shelf – it’ll need to be custom software, network aware (so the data can be fed back to the per-site data storage, allowing the devices themselves to be as ‘dumb’ as possible). Do you think $5,000,000 is cheap for that? Remember, the software would have to be aware of the full candidate list for every electorate they might be deployed to (even if only to simplify the logistics).
We’re not going to get into the fact that a good percentage of our population (let’s be kind, and say 5%) will struggle with the process, simply because it’s NOT pencil and paper, meaning that each location will need more people available to assist those who need it with the process. We’re not going to go into them in any depth, but be aware that this also raises privacy concerns – the way we vote is supposed to be private information.
So far, we’re spending $254,161,200. Add that to yesterday’s $102,975,600, and we get $357,136,800. And we still haven’t got the data to somewhere central for collation. Nor, for that matter, have we materially reduced any of the existing costs, which I’ve heard quoted as around $160 million. All we’ve really removed is the cost of pencils, paper, and printing, unlikely to total more than about $40,000,000.
Tomorrow: Why we didn’t network the whole process.